Development and validation of a three-dimensional deep learning-based system for assessing bowel preparation on colonoscopy video.

Background: The performance of existing image-based training models in evaluating bowel preparation on colonoscopy videos was relatively low, and only a few models used external data to prove their generalization. Therefore, this study attempted to develop a more precise and stable AI system for assessing bowel preparation of colonoscopy video. Methods: We proposed a system named ViENDO to assess the bowel preparation quality, including two CNNs. First, Information-Net was used to identify and filter out colonoscopy video frames unsuitable for Boston bowel preparation scale (BBPS) scoring. Second, BBPS-Net was trained and tested with 5,566 suitable short video clips through three-dimensional (3D) convolutional neural network (CNN) technology to detect BBPS-based insufficient bowel preparation. Then, ViENDO was applied to complete withdrawal colonoscopy videos from multiple centers to predict BBPS segment scores in clinical settings. We also conducted a human-machine contest to compare its performance with endoscopists. Results: In video clips, BBPS-Net for determining inadequate bowel preparation generated an area under the curve of up to 0.98 and accuracy of 95.2%. When applied to full-length withdrawal colonoscopy videos, ViENDO assessed bowel cleanliness with an accuracy of 93.8% in the internal test set and 91.7% in the external dataset. The human-machine contest demonstrated that the accuracy of ViENDO was slightly superior compared to most endoscopists, though no statistical significance was found. Conclusion: The 3D-CNN-based AI model showed good performance in evaluating full-length bowel preparation on colonoscopy video. It has the potential as a substitute for endoscopists to provide BBPS-based assessments during daily clinical practice.

Nanoscale Calculation of Proton-Induced DNA Damage Using a Chromatin Geometry Model with Geant4-DNA.

Monte Carlo simulations can quantify various types of DNA damage to evaluate the biological effects of ionizing radiation at the nanometer scale. This work presents a study simulating the DNA target response after proton irradiation. A chromatin fiber model and new physics constructors with the ELastic Scattering of Electrons and Positrons by neutral Atoms (ELSEPA) model were used to describe the DNA geometry and the physical stage of water radiolysis with the Geant4-DNA toolkit, respectively. Three key parameters (the energy threshold model for strand breaks, the physics model and the maximum distance to distinguish DSB clusters) of scoring DNA damage were studied to investigate the impact on the uncertainties of DNA damage. On the basis of comparison of our results with experimental data and published findings, we were able to accurately predict the yield of various types of DNA damage. Our results indicated that the difference in physics constructor can cause up to 56.4% in the DNA double-strand break (DSB) yields. The DSB yields were quite sensitive to the energy threshold for strand breaks (SB) and the maximum distance to classify the DSB clusters, which were even more than 100 times and four times than the default configurations, respectively.